Download presentation
Presentation is loading. Please wait.
Published byElijah Preston Modified over 9 years ago
1
Contiguous area cartograms Ingeborg Groeneweg
2
Introduction What are cartograms Difficulties creating cartograms History: previous approaches Current approach in-depth Summary
3
Cartograms Resizing regions of map by geographically related parameter Other way of representing the same: –3D-map –choropleth
4
Cartograms Area resizing: x y = h u v Shape preservation Topology preservation Contiguous
5
Example
6
2 parameter example
7
Difficulties Resizing area, preserving shape, preserving angle
8
Error measure Optimization problem Error measure –Area error –Shape error
9
History: Rubber map method Tobler, 1973 One of the first cartogram algorithms Used for population districting Idea: –put a dot for every person on a rubber map –Stretch the map until every dot is at equal distance Problem: –poor performance –Large area error –Overlapping shapes
10
Pseudo cartogram Tobler, 1986 Reduce area error Starting point for rubber map method
11
Rubber sheet distortion Dougenik, Chrisman, Niemeyer, 1985 Improvement on Tobler Difference: computing “force” on one polygon per iteration Overlapping shapes occur infrequently
12
DEMP Selvin et al.,1984 Density Equalized Map Projection (DEMP) Used to detect non-random distributions of disease Calculate spatial magnificent factor Radial transformation projected on selected area
13
Line integral Gusein-zade and Tikunov, 1995 Stokes theorem and line integrals
14
Forced-based Kocmoud and House, 1998 Alternately optimize shape and area error Superior to former methods
15
Cartodraw Keim, north, panse, 2004 Goal: –creating cartograms on the fly –Small error Cartodraw: –Simplify shape –Define error functions –Scanline –Main algorithm
16
Cartodraw: simplify shape Few vertices important for defining shape Vertices almost no noticeable difference: –angle near 180 degree –With short edges Different reduction algorithm for global shape and inner vertices
17
Global polygon reduction Only look at vertices v with d( v )> f( v ) –v at outer polygon –v do not belong tomultiple polygons Determine least important vertex w Finding polygon p where w is part of Counting difference d between p before and after removing w Remove w if d < constant
18
Least important vertex
19
Inner polygon reduction Remove all interior vertices v with d( v ) = 2 Reintroduce few vertices
20
Cartodraw Area error function Shape similarity function Scanline algorithm Main algorithm
21
Area error function Relative area error of polygon p j Area error for set of polygons P
22
Shape error function Translation, scale and partially rotation invariant Euclidean distance in Fourier space useful for shape similarity measure Use of differential geometric curvature of polygons Curvature will be square wave function
23
curvature
24
Example curvature
25
Fourier transformation Approximate function by summing sine and cosine Fourier approximation is defined as:
26
Fourier of square wave
27
Scanline Scanline sl = Line segment of arbitrary position and length Incrementally reposition vertices along scanline
28
Scanline Scaling factor
29
Cartodraw: main algorithm Transformation applied for each scanline –If E rel and shape distortion below certain threshold changes are retained Test improvement of area error
30
Automatic vs interactive Automatic generation of scanlines: –Fixed grid of horizontal and vertical scanlines –Resolution can be varied Interactive position of scanlines –Better results
31
Summary No ideal solution Several approaches reviewed Poor performance Handmade solutions superior
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.